11.5 Types of statistics disseminated#
This section covers the various types of data disseminated by an NSO. An NSO primarily disseminates macrodata although dissemination of geospatial data is becoming increasingly common. Microdata can also be disseminated, but this tends to be the exception.
11.5.1 Macrodata#
Macrodata is a term used to describe data generated by aggregating microdata according to statistical methodology. Examples of such aggregated data include unemployment statistics, demographics and GDP. There are few specific technical or legal issues relating to disseminating macrodata as they do not, in general, have confidentiality or data volume concerns.
11.5.2 Geospatial data#
An increasing number of NSOs are disseminating integrated geospatial information and statistics, including a geographic component. This means that the records in a geospatial dataset have implicit locational information such as Global Positioning System (GPS) (or geographic) coordinates. A geographic information system (GIS) is a system designed to capture, store, manipulate, analyse, manage, and present geospatial data (See. Chapter 9.4 — Geospatial data).
The challenges of disseminating integrated geospatial and statistical data mentioned in the GLOS guidance include:
“Official statistics typically disseminate statistical data free of charge according to open data principles. This may not always be the case for geospatial data provided by the national mapping agency or regional authorities. Therefore, access to geocoded data is not self-evident or could be costly for statistical authorities.
Statistical offices disseminate statistics in formats, such as tables, maps, graphs, infographics, news releases, public-use files, etc. The geospatial community uses traditional vector data and prepares large grids to disseminate their earth observations data. Publishing statistics on smaller areas and in an increasing number of formats makes the risk of disclosure higher. Very detailed aggregations pose risks to the privacy of individual data. Furthermore, the data protection methods used by mapping agencies and statistical offices differ, which can make the comparison of data difficult.”
11.5.3 Microdata#
Microdata generally contains information on individual persons, households or business entities collected through a survey or interview. In these datasets, each row typically represents an individual and each column an attribute such as age, gender or address. Microdata can also be made up of data on individuals collected from governmental administrative systems and registers. Microdata is used in official statistics for the production of aggregate information.
Microdata provides the underlying data for addressing critical development challenges such as poverty, gender inequality, and food insecurity. They can be aggregated into macrodata or, more commonly, statistical indicators, providing counts and averages at the level of a group, a country, or region. Because it contains disaggregated information about small population groups or areas lost in the process of compiling statistics indicators, microdata can provide a more nuanced, multidimensional view of the needs of vulnerable population groups that is essential for validating previous analyses, testing new hypotheses, and designing programs.
Some NSOs provide microdata for research and analysis purposes. It generally requires significant effort to produce high-quality microdata and the associated creation and documenting of microdata files, creating access tools and safeguards, and supporting and authorising inquiries made by the research community.
Dissemination of microdata in official statistics needs to be done carefully because it may lead to indirect or direct identification of the reporting unit, and measures must be taken to ensure anonymity. The privacy of data collected is core in protecting respondents, particularly for sensitive data set on health, gender violence or victimisation, thus ensuring the quality of and the trust in official statistics. Direct and indirect identification of individuals should be impossible with microdata publication while for research purposes, indirect identification could be possible and therefore, such accesses strictly regulated (see Chapter 15.2.17 — Data security).
NSOs often charge researchers for such access to offset the extra costs of making the microdata available. There are also costs associated with the checking of outputs to ensure confidentiality.
For more information, see also Annex 5 - Why Share Microdata? – A View from ODW.
Eurostat microdata reference guide (🔗).
OECD expert group for collaboration on microdata access (🔗).
FAO microdata catalogue (FAM) (🔗).
Guidelines and good practices for managing statistical confidentiality and microdata (🔗).
INSEE France Statistical confidentiality and data protection (🔗).
UN Principles and Guidelines for Managing Statistical Confidentiality and Microdata Access (🔗).